Mathematics > Classical Analysis and ODEs
[Submitted on 18 Jun 2024 (v1), last revised 21 Aug 2024 (this version, v2)]
Title:The Laplace asymptotic expansion in high dimensions
View PDF HTML (experimental)Abstract:We prove that the classical Laplace asymptotic expansion of $\int_{\mathbb R^d} g(x)e^{-nz(x)}dx$, $n\gg1$ extends to the high-dimensional regime in which $d$ may grow large with $n$. We formulate simple conditions on the growth of the derivatives of $g$ and $z$ near the minimizer of $z$ under which the terms of the expansion and the remainder are bounded as powers of $(\tau d/\sqrt n)^2$. The parameter $\tau$ controls the growth rates of the derivatives and can be potentially large, but we obtain a useful expansion whenever $\tau d/\sqrt n\ll 1$. Our result relies on a new and transparent proof of the Laplace expansion valid for any $d$. The proof leads to a new representation of the terms and remainder, which is crucial in obtaining tight control on the growth of these quantities with $d$. To demonstrate that our bounds are tight, we consider the case of a quartic $z$ and $g\equiv1$, showing that the $k$th term is precisely of the order $(d^2/n)^k$ for all $k$, and that our bounds on the terms are of the same order of magnitude. We also apply our results to derive a high-dimensional Laplace expansion for a particular function $z$ arising in the context of Bayesian inference, which is both random and depends on $n$. We show that with high probability, the terms of the expansion and the remainder are bounded in powers of $d^2/n$. In both of these examples, $\tau=1$.
Submission history
From: Anya Katsevich [view email][v1] Tue, 18 Jun 2024 15:19:02 UTC (27 KB)
[v2] Wed, 21 Aug 2024 01:51:55 UTC (64 KB)
Current browse context:
math.CA
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Connected Papers (What is Connected Papers?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
alphaXiv (What is alphaXiv?)
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Hugging Face (What is Huggingface?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.